Quantifying Uncertainties in Wind Atlases

English Abstract:

The Great Lakes present a large potential for offshore wind development due to the abundant wind resource and the proximity to large population centers. With low water depths, Lake Erie has been a focus of wind energy projects for the past few years. In this study, observed winds from 12 coastal stations, 4 buoys, and ESA’s Synthetic Aperture Radar during 2002-2012 are used to generate an observational wind atlas for Lake Erie. The in situ data provide temporally consistent data at various heights and at point locations. The satellite data represent equivalent neutral winds at 10 m, provide spatial information at a relatively high resolution but are temporally sparse. Each data set is treated separately to take advantage of their respective strengths and address their weaknesses. A methodology is presented to correct for the consistent temporal gaps introduced in offshore observations when the buoys are removed from the lake and the satellite platforms cannot estimate the wind stress due to ice formation. A method of ratios is applied on the mean wind and power density to correct the buoy generalized wind climates. The wind class method is used to address the disjunction in the available satellite scenes. All processing steps reduce the uncertainty in the results. A method is presented that integrates the individual wind climates into a wind atlas for the region, under the common assumption of neutral stratification offshore. The final area-averaged RMSE is 0.1 m/s and the mean predicted resource is < 314 W/m2 at 90 m in the center of the Lake. To test the neutral stratification assumption, observations over Lake Erie were obtained during a one-month measurement campaign from a buoy, a sonic anemometer, a profiler lidar and a scanning lidar. These data are used to investigate in detail the local characteristics of the wind flow and quantify the impact of stability on the wind climate of the region and the uncertainty in the wind atlas estimation.